444 Lecture 13
2024-03-05
In a formal representation of a game we specify:
Most of the games we’re going to look at have the following characteristic.
There are two natural ways to specify the outcome of a game.
It is, more or less, desirability.
| C | D | |
|---|---|---|
| C | 3,3 | 0,5 |
| D | 5,0 | 1,1 |
| C | D | |
|---|---|---|
| C | 3,3 | 0,5 |
| D | 5,0 | 1,1 |
The big tension:
What real life situations might be like this? I.e., which real life situations have these features:
Were you able to register:
I’ll change the numbers to allow the game to start.
You’ll be randomly paired with someone else in the room, and you’ll play Prisoners’ Dilemma 5 times. After each play, you’ll see what the other person did the previous round.
Your aim is to get as many points (or ‘dollars’) as possible over the five rounds.
You’ll be randomly assigned Row or Column; the game is symmetric so this doesn’t make a different to strategy.
| C | D | |
|---|---|---|
| C | 20,20 | 0,20 |
| D | 15,0 | 15,15 |
Go back into veconlab, and you should be set up with a different person from the PD rounds to play a single round of Stag Hunt.
| C | D | |
|---|---|---|
| C | 0,0 | 0,1 |
| D | 1,0 | -20,-20 |
Go back into veconlab, and you should be set up with a different person again from the last two rounds to play a 5 rounds of Hawk-Dove.
Lessons from Iterated Prisoners’ Dilemma